Hey folks, it’s Jeff Crusey, Per Aspera cofounder and your resident VC on the team (which means I’m an expert on every single topic that exists 😂).
On a more serious note, lately, I’ve been fixated on a very specific topic: Why is the U.S. spending tens of billions to acquire the very same resource we’re throwing away every day?
In today’s A1 story, I unpack this question and make the case for a strategic recycling program for critical minerals, with my editor Ryan riding shotgun.
IN THIS WEEK’S EDITION:
💎 “One Man’s Trash Is Another’s Critical Feedstock”
📬 Now, for something new…introducing PA’s Question Bank
🗞️ An American succes story, a Moon-pilled SpaceX, & Intel’s latest
Forwarded this? Subscribe to Per Aspera here.
Last week, the White House announced Project Vault, a $12B program to build a strategic reserve of dozens of critical minerals as a hedge against Chinese supply disruption. This is the most aggressive domestic minerals play in a generation, and a big step in the right direction.
And yet, we can't shake the feeling we're overlooking a high-leverage opportunity from the unlikeliest of places: our trash cans. Across the press releases and procurement plans we’ve read, recycling is still a small, scattered line item when you size it up against the big, shiny stockpile.
So, today’s question: why is recycling a rounding error in our strategy to break the mineral chokehold?
The Case for Recycling
This February, Americans will dump ~500,000 tons of dead electronics. Today alone, we’ll send 416,000 phones, 142,000 computers, and 37,000 cars to landfills, shredders, or low-road export channels. These devices being put to digital pasture are laden with critical minerals and rare earth elements (REEs) — already mined, already refined, already in the alloys our government, defense industrial base, and critical technology industries are so desperate to secure.
Conservatively, a third of U.S. e-waste still ends up landfilled or burned, and somewhere between a tenth and two-fifths of ‘recycled’ devices quietly leave the country. All in, recovered REEs from end‑of‑life products cover ~1% of annual U.S. demand.
As we put up tens of billions to dig rock and build vaults, we are spending an order less on recycling. There’s no silver bullet here. Mines, refineries, science, finance, and alliances all must play their part. But we think recycling could play a bigger part in closing the strategic gap.
Today we share eight reasons why:
The single-use problem. You know what they say: One Man’s Trash Is Another Man’s Concentrated Feedstock. A neodymium magnet from a hard drive is ~30% REE by weight; typical ore runs below 2%. That’s a 15× enrichment factor before you even touch a chemical. This is the crux of the issue: we pay to import critical materials embedded in finished goods, then bury/export them at end-of-life when there’s still residual value.
Weaning off Chinese refineries. Cocktail party chatter on this matter (unclear if this has ever actually happened) conflates mining and refining. But as you, dear reader, of course already know, our chokepoint is midstream processing. China holds ~67% of global mining (big but manageable), 91% of refining, 87% of oxide separation, and ~92% of magnet production. The good news: a recycled magnet is already a finished alloy, not mixed ore. Salvage more from magnets, motors, and battery materials → re-enter the chain after the most chemically complex, pollutive, China-dominated step.
Heavy REE independence. Whereas light REEs can be mined in California or Allied countries, the heavies largely originate from ionic clay deposits in southern China and Myanmar. But! Good news again. We’ve been importing heavy REEs for decades in our finished products. Dysprosium and terbium, two of the heavy REEs that folks talk about most at these fictional cocktail parties, let permanent magnets perform at ultra-high temperatures in jet engines, missile guidance systems, and submarine drive motors. They can be found in decent quantities across spent hard drives, MRI machines, EV motors, and wind turbines. Beyond the stockpiles we’re building now, our domestic e-waste stream may be the only scalable near-term source of defense-critical heavy REEs that doesn't require Beijing’s stamp of approval.
Keep reading for the next five reasons — yes, the very next one gets into unit economics — and then drop us a line letting us know what you thought.,
FROM THE DESK OF DAN GOLDIN….
Goldin popping in! Sobriety is a big part of the Per Aspera value system, so I (and we) tend to gravitate to the “doom and gloom” realities of the American Renaissance. But today, I wanted to pop in with an American success story that brings me hope!
A few years ago, our beloved Micron had ~5% of the high-bandwidth memory (HBM) market — the specialized memory chips that sit right next to AI processors — while SK Hynix and Samsung controlled the remaining ~95%.
Now Micron is ramping toward the 20-25% range within the next year or so, according to industry forecasts and management’s targets.
They got there by solving hard technical problems. Micron's 12-layer HBM3E, which NVIDIA qualified for Blackwell systems, delivers about 20% lower power than competing 8-high parts (some analysts estimate even bigger gains!).
HBM supply was locked up between two Korean companies. Now, thanks to Micron, we’ve wedged ourselves into the market as a serious American alternative. A U.S. hardware company catapulting from low-single-digit share toward roughly one-fifth of this strategic market in just a few years. That is proof of the American Renaissance at work!
I love all of you doing hard things ❤.
SPACEX IS MOON-PILLED:
"SpaceX has already shifted focus to building a self-growing city on the Moon, as we can potentially achieve that in less than 10 years… the overriding priority is securing the future of civilization and the Moon is faster."
After a decade of “straight shot to Mars” rhetoric, SpaceX has retired the bit and committed to colonizing the lunar surface first. Two reasons why, as we see it:
The simple practicality of it. You can launch to the Moon every 10 days (vs. every 26 months) and it’s only a two-day trip (vs. half a year). This lets you run the build-fly-break-fix loop on an Earth-adjacent product cycle, not a multi-year, unite-the-world-and-rescue-Matt Damon cadence.
Second, and this is not something you hear a lot about the Moon, but SpaceX seems to be rapidly building conviction around a real lunar business case. Musk’s been talking up lunar mass drivers a lot lately: electromagnetic launch railroads that could yeet refined metals, solar wafers, and space-built parts into orbit for pennies on the kilogram.
The Moon is so back. With Artemis II on deck, Blue Origin’s Mk1 lander gearing up for the first Blue Moon cargo run, China pressing ahead on ISRU, and Austin’s* Own (*Cedar Park) — and Per Aspera Founding Sponsor — Firefly having clinched the honor of the first fully private U.S. cargo delivery to the lunar surface…we’re thinking we may need to drop a deep-dive Antimemo on the Moon real soon.
AND….QUOTE OF THE WEEK:
“Almost every CEO, they call me and say: ‘Lip-Bu, can I have more? I’m your friend. I’m your customer, the most important customer, I want to have more of that.’ So I think it’s encouraging for me to see that compute has become very important.”
That’s Intel CEO Lip-Bu Tan, fresh off an earnings call during which he effectively distilled the chipmaker’s assignments: be everyone’s foundry and sell every CPU you can make. Tan is probably pretty pleased with the “CPUs-are-back” narratives forming around AI right now, with new enthusiasm for the processor across datacenter pipelines, small/edge/on-device inference models, and “AI PC” marketing. But Tan also had a warning: China’s chipmaking ecosystem is rapidly closing the gap with the U.S. across design tools, equipment, and manufacturing, and running laps around us on power deployments. And Intel doesn’t have the luxury of resting on its laurels; it still has real ground to make up to get to parity with the Asian fabs. (ICYMI: read our take on why Intel’s last mission should be to transform itself from a conflicted conglomerate into America’s sovereign foundry.)
Per Aspera Friends in High, Hard Places
Skyryse has raised a $300M+ Series C, vaulting it into unicorn territory (~$1.15B valuation). The aerial autonomy company is focused on completing FAA certification and scaling its SkyOS (tagline: “One System, Any Aircraft) across new aircraft. We will forever be impressed with how Skyryse took on and conquered helicopters, the final boss of flight automation. Read more here.
Stoke Space has extended its previously announced Series D to a total of $860M. The mission remains the same: a fully, rapidly reusable medium-lift launch vehicle. Stoke will continue engine/vehicle testing, infrastructure buildout, and the forward march toward a system that can be turned around like an aircraft. Congrats to Andy and everyone else on Team Stoke!
Jared, if this finds its way into your inbox between meetings, sorties, hiring, and budget reviews — HBD! You’ve made the Administrator’s chair look fun, exhilarating, and deadly serious all at once, and your impact already shows as a clear upward inflection in NASA’s trajectory. Thanks for taking on The Hard Stuff. Here’s to another lap around the sun!
The votes are in…
Last week, we asked: ‘What’s the biggest matchup you care most about?’
US vs. China critical technology ecosystems took the plurality at 41%, with SpaceX vs. Everyone Else close behind at 33% and Atoms vs. Bits at 22%. Some readers went the extra mile to write in, and we wanted to share some of the thought-provoking responses:
Voted U.S. vs. China: “China’s ability to strategize over the long term and build supply chains, ecosystems, and synergies around future technologies is a winning strategy. If the U.S. continues to think in the short term and prop up last-century technologies…exciting work will move outside the U.S.” A longstanding theme we always find ourselves revisiting: how can America match long-horizon industrial strategy with its short-term cycles?
Voted Bits vs. Atoms: “Embodied AI was extremely buzzy this year at CES; let's see where this goes!” Firstly, we agree, and second, while bits vs. atoms of course isn’t a zero-sum game, we’ll always have the backs of all of you building hard things in the real world!
Voted U.S. vs. China: “It seems like a post-Sputnik era again.” We’ll just let this one sit with folks…
|